TPS56637: Power supply chip in the application process of dc-dc chip damage short-circuit phenomenon.

Hi Zhonghui,

According to the problem description, after removing TPS56637 chip, the positive and negative terminal (P31 1& 2 connecter) were shorted, if this is true, I highly suspected the input cap C1/C3/C4 may be shorted or mostly shorted. Can you double check the input cap status or replace new one? What's more, what is the input cap voltage rating, 25V/35V/...?

Thanks,

Lishuang

"we removed the TPS56637 chip and found that the positive and negative terminals were shorted." This means the pins of  the  TPS56637 chip were shorted,the other elements On the pcb were ok.

Hi Lishuang,

Min is my customer and the initiator of this problem. He knows the details very well.You can communicate with him directly.

Thanks.

Best regards,

Hi Min,

1. Have you done ABA swap test to confirm the unit was damaged? Seems you only did half of ABA swap test. Please also solder the 'damaged' chip onto a good board, to see whether the issue will happen.

2. From your description, the chip's Vin pin was short to ground? Could you measure all pins' impedance (to ground) of good chip v.s. damaged chip? Any other pins short to ground?

3. I found you use 2x22uF as Cout. Have you checked the 22uF cap derating ratio at 12V? Please make sure the actual Cout value is over 25uF, as suggested in datasheet Table 4. 

Hi Min,

Sorry, I misunderstood the positive and negative terminals are for board.

Adding one comment based on Miranda's,

4. Which schematic or which inductor is the damaged chip used? L1 or L2? I found that L1 saturated current is only 3.5A while your application is 6A. If the L1 is saturated, the chip may be damaged. Would you mind to change a new inductor with higher current rating, higher than IOC max 8.6A would be good.

Thanks,

Lishuang

1. After we removed the tps56637 from the board, and used the multimeter to measure the positive and the negative pads on the board, it was not shorted, it was ok. Also measured the pins of the TPS56637( PIN8 AND PIN9) that we removed from the board, the PIN8(VIN) and PIN9(PGND) was shorted. After done these, we thought the chip was broken, so we didn't solder the “broken chip” on the good board.
2. We measure the pin9(PGND) with other pins of the broken chip, the pin9 with these pins{pin3(AGND), PIN5(NC), PIN6(SW), PIN7(BOOT),PIN8(VIN)} are shorted.
3. According to your suggestion, we will test with capacitors greater than 25uF.

Thanks

MIN LI

Both circuits in the schematic have been tested, and both chips are broken. The reason why the rated current of the inductors is different is that Party A's project needs to verify whether the two inductors have the same characteristics and can be replaced (regardless of the current size).

Thanks

MIN LI

Hi Min,

Okay. 

1. Can we understand that both circuits layout/components are same, except inductor? Both of them are damaged.

2. When does the damage happen? during power on or long time under 6A steady state?

3. Figure 1 shows it can be stable at 24Vin-12Vout-6A, can you help to measure the VIN SW Vout and inductor current waveforms during power on and steady state? If both of the new device with SCH1 and SCH2 can be only normal at 1A, would you mind to measure above waveforms under 1A under whistling?

4. I will double check the layout later. Now, our Allegro license server has some error, can't open the layout now.

Thanks,

Lishuang

According to your proposal, we added two 10uf aluminum electrolytic capacitors yesterday at the output end, and now the test chip is not broken. In turn, the output current was 1a,2a,3a,4a,5a,6a, and now the chip can be run properly and the chip is not broken. But there's a new little problem, which is that after the 6a long run, for example, the output voltage that runs 20 seconds will immediately fall and it will recover immediately(We recorded a video about this process), and we used the thermocouple thermometer to test the temperature of the chip shell at about 116 degrees Celsius. So we wonder if it is the internal heat of the chip triggers the heat protection in the inside of the chip? If that's the case, is there some good suggestion on Heat Dissipation about this chip.

According to your proposal, we added two 10uF/50V aluminum electrolytic capacitors yesterday at the output end, and now the test chip is not broken. In turn, the output current was 1a,2a,3a,4a,5a,6a, and now the chip can be run properly and the chip is not broken. But there's a new little problem, which is that after the 6a long run, for example, the output voltage that runs 20 seconds will immediately fall and it will recover immediately (We recorded a video about this process, this phenomenon occurred at 21s), and we used the thermocouple thermometer to test the temperature of the chip shell at about 116 degrees Celsius. So we wonder if it is the internal heat of the chip triggers the heat protection in the inside of the chip ? If that's the case, is there some good suggestion on Heat Dissipation about this chip.